Measuring the distribution of wages in the United States from 1996 through 2010 using the Occupational Employment Survey

By nearly every measure, the distribution of wages in the United States has been growing more unequal since the late 1970s.¹ In this article, we show how data from the Occupational Employment Statistics (OES) survey can be used to measure the changing distribution of wages in the United States since the late 1990s, and we describe aspects of this changing distribution of wages that the OES data are uniquely well suited to describe.

There is an enormous literature devoted to understanding the nature and sources of growing wage inequality. Much of this literature relies on data on the wages and characteristics of individual workers from the Current Population Survey (CPS). For example, many studies have examined the changing composition of the workforce and changing returns to education and experience² and the growing dispersion of wages among the most educated and experienced workers.³ Growing inequality has been attributed to the differential impact of technology on differing portions of the worker skill distribution,⁴ to the differential ‘offshorability’ of occupational tasks,⁵ to changing labor market institutions such as declining unionization levels,⁶ to the declining real value of the minimum wage,⁷ and to the growing fraction of workers subject to performance-based pay from their employers.⁸ Although these explanations for growing inequality are concerned with the policies and incentives faced by employers, the CPS data used in this literature contain little if any information on the businesses employing these workers.

Other researchers have studied wage inequality using data collected from employers. These studies have built on evidence that establishments play an important role in determining individual wages.⁹ Several authors have used employer microdata to study growing earnings variance in the United States from the mid-1970s to the early 2000s, and have found that the increasing variance is due more to variation between establishments than to variation within establishments.¹⁰ However, the employer-provided data used in this literature have limited information on the characteristics of workers in each establishment.

This disconnect between wage inequality studies based on individual worker microdata versus studies based on employer-provided microdata can be bridged with data from the Occupational Employment Statistics program. The OES data are unique in containing information on establishment characteristics as well as information on wages and occupations for all employees within surveyed establishments. Our analysis of the occupational data in the OES complements a small yet growing literature that analyzes the relationship between occupations and increasing wage inequality.¹¹

The OES data are collected from a large semiannual survey of establishments. These data allow us to use a single source of wage information as we decompose increasing wage inequality in the United States into its within- and between-occupation components and into its within- and between-establishment components. However, the OES was not designed to study the time series of wage inequality (or any other time series¹²), and as such, we describe the adjustments we have made to the confidential OES microdata that are necessary for analyzing wage variance trends.

Using the OES data, we replicate many of the wage variance trends that other authors have found using the CPS data. We show that most of the growth in wage inequality for the period of our study has occurred within the private sector, within particular industry groups such as professional and business services, and within occupational groups such as healthcare. We show that occupations explain a growing fraction of all wage variance, and, in an examination not possible with the CPS data, we find that within the private sector, wage differences among establishments explain far more of the level and the trend of wage variance than do wage differences among occupations.

Using the OES data to make comparisons over time

The Occupational Employment Statistics survey is designed to measure occupational employment and wages in the United States by geography and industry and is the only such survey of its size and scope. Since 1997, the OES has covered all workers in the United States except for agricultural workers, private household workers, and unincorporated self-employed workers without employees. Every year, approximately 400,000 private and local government establishments are asked to report the number of employees in each occupation within specific wage intervals. At the same time, complete employment data are obtained from the federal and state governments.¹³ An example of a portion of a survey form is given in the following exhibit.

For large establishments, the survey form lists 50 to 225 detailed occupations (occupations preprinted on the survey form are selected based on the industry and the size of the establishment). Small establishments receive a blank survey form and write in descriptions of the work done by their employees. These employer-provided descriptions are coded into occupations by staff in state labor agencies (as part of the OES federal–state partnership). Wage intervals on the OES survey form are given in both hourly and annual nominal dollars.¹⁴ To calculate average wages, the OES program obtains the mean of each wage interval every year from the National Compensation Survey (NCS). These mean wages are then assigned to all employees in that wage interval.¹⁵ Approximately 25 percent of jobs in the OES data are imputed (with an establishment nonresponse rate of approximately 20 percent).

The OES program underwent a large number of improvements between 1996 and 2002, many affecting the comparability of data over time. The first year of a truly national OES survey was 1996, when wage data were collected from 400,000 establishments in every industry and every state. In 1997, the program began including establishments with less than 5 employees in the sample. Improvements to the sampling and weighting methodologies took place in 1998. In 1999, changes in the width of the wage intervals were instituted, making them more uniform in natural logarithm (wage) terms, as well as the addition of a 12th wage category. Nominal wage intervals printed on the form were shifted again in 2005 and 2009. In addition, 1999 was the first year that the occupations used in the OES were those of the Standard Occupational Classification system. The year 2001 was the last in which data were collected in October, November, and December; beginning with November 2002, data have been collected from 200,000 establishments each November and 200,000 establishments each May. The 2002 sample was the first designed using the North American Industry Classification System (NAICS) in place of the older Standard Industrial Classification (SIC) system. Each of these survey changes is designed to improve the quality of OES data but has the downside of creating difficulty in comparing estimates from year to year.

The OES survey is not designed to produce time-series statistics. To reduce variance and include data in each estimate from large employers that are surveyed only once each 3 years, published estimates from the OES program are based on the previous 3 years of data. Over each 3-year survey cycle, large establishments are sampled with certainty, and no establishment is sampled more than once. Before any of the estimates described in this paper were made, much preparatory work was devoted to the creation of appropriate weights in order to have the OES data in each individual panel be self-representing. Using the methodology described in the 2009 chapter entitled “Are the new jobs good jobs?”¹⁶ we reweight the data to November (or May) benchmarks of total employment by detailed industry and by broad industry and establishment-size groups from the Quarterly Census of Employment and Wages (QCEW). This reweighting forces the establishments in each separate panel to match the overall distribution of establishments during that November (or May), by detailed industries and by size groups within broad industries. This reweighting performs well for national-level estimates for broad categories of industries and occupations but would be inappropriate to use for more detailed levels of geography, industry, or occupation.¹⁷

We have assembled OES microdata for the private sector from fall 1996 through May 2010 and for both the private and government sectors from fall 1998 through May 2010.¹⁸ After making the adjustments to weights described above, we create time series using each survey date.

Comparing OES data to CPS data

Because OES data have never before been used to study national wage inequality trends in this way,¹⁹ we show how measures of inequality derived from OES data compare with more familiar measures derived from the CPS. Our preparation of CPS microdata follows the methodology of Thomas Lemieux.²⁰ Although the CPS-based literature on wage inequality uses several measures of wage inequality, including the gaps in wages between the 90th, 50th, and 10th percentiles of the wage distribution as well as the overall variance of wages, we examine only the variance of wages. This is because the OES data are collected in intervals (as described below), so we cannot calculate exact percentiles of the OES wage distribution without making assumptions about the distribution of wages within each interval.²¹

Following Lemieux, we use the CPS Outgoing Rotation Group (ORG) samples. In the CPS data, we define hourly wages as the reported hourly wage for those who are paid by the hour (approximately two-thirds of employed people), and as the ratio of usual weekly earnings divided by usual weekly hours for those not paid hourly. We use Lemieux’s imputation of usual hours for those who report that their usual hours vary. We multiply both hourly and weekly topcoded earnings by 1.4. For this project, we trim all hourly earnings in the CPS by dropping earnings for those reporting real hourly earnings less than $1 or greater than $100 (in constant November 2007 dollars). We also drop all imputed wage observations. Trimming the very few outliers in the hourly wage distribution has a very substantial effect on the time series of variance. Removing imputations slightly raises the variance in any given year, but the time trends of the variance of trimmed wages are similar whether imputed values are or are not included in the data. To be consistent with the manner in which wages are treated in the OES, we raise all subminimum hourly wages to state-specific minimum wage levels. This reduces the overall variance of wages in the CPS with little impact on the time trend of the variance in wage levels.

There are two important differences between the scope of OES and CPS earnings data that we are unable to control for. First, CPS earnings data are based on an individual’s main job (individuals with multiple jobs are only asked about their earnings on their main job), whereas the OES earnings data are recorded for all jobs. Second, earnings questions in the CPS are not asked of the self-employed (either incorporated or unincorporated), while the OES does include wages for the incorporated self-employed.

The measurement of wages is very different in the CPS and OES. The CPS collects wage information in exact dollars and cents, while the OES survey collects wage information in discrete wage intervals (11 wage intervals for 1996–1998, and 12 intervals thereafter). To address the importance of this difference, we assign each observation in the CPS microdata to the OES wage interval in which that wage falls. Figure 1 shows the overall distribution of employment in the OES and the CPS by these wage intervals, dropping observations with imputed wages, for the fall of each calendar year.²² For example, looking at the third panel in figure 1 titled “Fall 1998 all sectors,” we see that roughly 18 percent of jobs in both the CPS and the OES data are in the first wage interval, about 14 percent of jobs in both datasets are in the second wage interval, and so forth.

In 1998 and in every subsequent year, the distribution of employment by wage interval is quite similar between the OES and the CPS data. However, the distributions in 1996 and 1997 are markedly different in the first two intervals, although the distributions are essentially the same for all other intervals. We have no simple explanation for this discrepancy.²³ Because the distributions of wages between the CPS and the OES are so different for 1996 and 1997, we do not use these years in our analysis of the distribution of wages. The remainder of this paper shows results based on data beginning in 1998, when the distribution of employment by wage interval is quite similar between the OES and the CPS data.

There are other notable features of the wage distributions in figure 1. First, the changes in nominal wage intervals that took place in 1999, 2005, and 2009 are evident in the figure. The intervals of various widths (in log terms) in 1996–1998 were replaced with intervals of approximately equal width (in log terms) in 1999. In 2005 and 2009, the endpoints of all the intervals were shifted slightly to keep pace with nominal wage growth. For example, in 1999–2004, there were six intervals for workers earning less than $20 per hour, but beginning in 2005, there were only five such intervals. Second, the interaction between these nominal wage intervals and inflation changes the shape of the distribution of “intervalized” wages (that is, wages expressed in discrete wage intervals) from period to period: the nominal wage distribution shifts to the right between years when the nominal wage intervals are unchanging but nominal wage inflation takes place.

View Chart Data

Figure 2. Overall employment levels in the CPS and OES, 1998–2010

Data source	Date	Employment
OES	Nov 1998	123,294,897
OES	Nov 1999	125,480,740
OES	Nov 2000	129,300,195
OES	Nov 2001	127,738,844
OES	Nov 2002	127,515,741
OES	May 2003	126,708,316
OES	Nov 2003	127,208,097
OES	May 2004	128,144,388
OES	Nov 2004	129,926,191
OES	May 2005	130,879,045
OES	Nov 2005	131,863,665
OES	May 2006	132,836,799
OES	Nov 2006	133,375,261
OES	May 2007	134,238,904
OES	Nov 2007	135,116,541
OES	May 2008	134,608,928
OES	Nov 2008	132,489,908
OES	May 2009	127,820,779
OES	Nov 2009	126,847,848
OES	May 2010	126,948,875
CPS	Jan 1998	123,427,052
CPS	Feb 1998	124,022,459
CPS	Mar 1998	124,719,704
CPS	Apr 1998	124,116,422
CPS	May 1998	125,571,176
CPS	Jun 1998	125,657,761
CPS	Jul 1998	125,700,428
CPS	Aug 1998	125,003,304
CPS	Sep 1998	125,635,001
CPS	Oct 1998	126,184,564
CPS	Nov 1998	126,698,304
CPS	Dec 1998	126,675,348
CPS	Jan 1999	127,074,364
CPS	Feb 1999	126,547,134
CPS	Mar 1999	127,215,844
CPS	Apr 1999	126,283,639
CPS	May 1999	128,262,064
CPS	Jun 1999	127,634,909
CPS	Jul 1999	127,860,510
CPS	Aug 1999	127,659,319
CPS	Sep 1999	127,759,093
CPS	Oct 1999	128,686,140
CPS	Nov 1999	128,844,355
CPS	Dec 1999	130,034,212
CPS	Jan 2000	128,413,186
CPS	Feb 2000	128,525,759
CPS	Mar 2000	128,729,404
CPS	Apr 2000	129,720,349
CPS	May 2000	129,717,783
CPS	Jun 2000	129,870,141
CPS	Jul 2000	130,000,971
CPS	Aug 2000	129,454,882
CPS	Sep 2000	129,524,293
CPS	Oct 2000	130,406,518
CPS	Nov 2000	130,691,246
CPS	Dec 2000	131,230,070
CPS	Jan 2001	129,347,907
CPS	Feb 2001	130,476,829
CPS	Mar 2001	130,104,877
CPS	Apr 2001	129,175,032
CPS	May 2001	130,253,933
CPS	Jun 2001	130,412,124
CPS	Jul 2001	130,668,585
CPS	Aug 2001	129,534,058
CPS	Sep 2001	130,010,085
CPS	Oct 2001	129,042,466
CPS	Nov 2001	128,852,968
CPS	Dec 2001	129,252,107
CPS	Jan 2002	127,415,223
CPS	Feb 2002	129,299,813
CPS	Mar 2002	128,930,318
CPS	Apr 2002	128,757,420
CPS	May 2002	129,635,450
CPS	Jun 2002	129,486,127
CPS	Jul 2002	130,036,022
CPS	Aug 2002	129,995,318
CPS	Sep 2002	130,246,681
CPS	Oct 2002	129,743,633
CPS	Nov 2002	128,711,667
CPS	Dec 2002	129,432,940
CPS	Jan 2003	127,795,722
CPS	Feb 2003	129,094,135
CPS	Mar 2003	128,857,618
CPS	Apr 2003	129,187,299
CPS	May 2003	130,044,238
CPS	Jun 2003	129,570,837
CPS	Jul 2003	130,103,174
CPS	Aug 2003	128,941,178
CPS	Sep 2003	128,860,560
CPS	Oct 2003	130,482,265
CPS	Nov 2003	129,337,513
CPS	Dec 2003	130,423,178
CPS	Jan 2004	128,437,954
CPS	Feb 2004	129,668,985
CPS	Mar 2004	130,263,102
CPS	Apr 2004	130,026,535
CPS	May 2004	130,368,884
CPS	Jun 2004	131,087,346
CPS	Jul 2004	132,070,049
CPS	Aug 2004	131,654,896
CPS	Sep 2004	131,468,797
CPS	Oct 2004	132,061,525
CPS	Nov 2004	131,776,983
CPS	Dec 2004	133,213,819
CPS	Jan 2005	130,546,689
CPS	Feb 2005	130,992,564
CPS	Mar 2005	131,856,157
CPS	Apr 2005	131,877,117
CPS	May 2005	132,947,991
CPS	Jun 2005	133,896,906
CPS	Jul 2005	134,096,260
CPS	Aug 2005	134,703,710
CPS	Sep 2005	133,956,152
CPS	Oct 2005	134,667,528
CPS	Nov 2005	134,701,175
CPS	Dec 2005	135,292,022
CPS	Jan 2006	132,751,244
CPS	Feb 2006	134,145,401
CPS	Mar 2006	135,046,306
CPS	Apr 2006	134,355,700
CPS	May 2006	135,473,702
CPS	Jun 2006	136,262,123
CPS	Jul 2006	136,600,925
CPS	Aug 2006	135,648,099
CPS	Sep 2006	136,280,054
CPS	Oct 2006	137,475,933
CPS	Nov 2006	137,967,128
CPS	Dec 2006	137,828,263
CPS	Jan 2007	136,067,560
CPS	Feb 2007	136,325,871
CPS	Mar 2007	136,806,452
CPS	Apr 2007	136,798,293
CPS	May 2007	136,597,325
CPS	Jun 2007	137,236,176
CPS	Jul 2007	138,634,060
CPS	Aug 2007	137,350,075
CPS	Sep 2007	137,205,316
CPS	Oct 2007	138,188,718
CPS	Nov 2007	138,800,136
CPS	Dec 2007	138,554,285
CPS	Jan 2008	137,491,096
CPS	Feb 2008	136,589,837
CPS	Mar 2008	137,298,262
CPS	Apr 2008	137,587,791
CPS	May 2008	138,026,323
CPS	Jun 2008	138,258,588
CPS	Jul 2008	137,772,583
CPS	Aug 2008	137,376,008
CPS	Sep 2008	136,729,666
CPS	Oct 2008	138,189,192
CPS	Nov 2008	137,414,683
CPS	Dec 2008	136,412,069
CPS	Jan 2009	133,864,078
CPS	Feb 2009	133,604,886
CPS	Mar 2009	132,736,147
CPS	Apr 2009	133,443,079
CPS	May 2009	132,998,055
CPS	Jun 2009	132,734,330
CPS	Jul 2009	133,350,354
CPS	Aug 2009	131,712,361
CPS	Sep 2009	130,605,602
CPS	Oct 2009	132,287,506
CPS	Nov 2009	131,830,333
CPS	Dec 2009	130,985,639
CPS	Jan 2010	129,949,844
CPS	Feb 2010	130,518,980
CPS	Mar 2010	130,483,843
CPS	Apr 2010	132,219,648
CPS	May 2010	131,809,284
CPS	Jun 2010	132,121,378
CPS	Jul 2010	131,927,866
CPS	Aug 2010	132,249,541
CPS	Sep 2010	132,224,975
CPS	Oct 2010	132,399,896
CPS	Nov 2010	131,881,036
CPS	Dec 2010	132,547,535

View Chart Data

 

Figure 3. Overall wage levels in the CPS and OES, 1998–2010

Data source	Date	Log real hourly wage	Log real hourly wage—no imputations
OES	Nov 1998	2.77	2.76
OES	Nov 1999	2.79	2.77
OES	Nov 2000	2.8	2.78
OES	Nov 2001	2.82	2.81
OES	Nov 2002	2.84	2.83
OES	May 2003	2.84	2.82
OES	Nov 2003	2.83	2.82
OES	May 2004	2.83	2.81
OES	Nov 2004	2.82	2.81
OES	May 2005	2.82	2.81
OES	Nov 2005	2.84	2.83
OES	May 2006	2.82	2.8
OES	Nov 2006	2.83	2.81
OES	May 2007	2.82	2.8
OES	Nov 2007	2.82	2.81
OES	May 2008	2.82	2.8
OES	Nov 2008	2.87	2.86
OES	May 2009	2.87	2.85
OES	Nov 2009	2.85	2.83
OES	May 2010	2.85	2.83
CPS	Jan 1998	2.7	2.7
CPS	Feb 1998	2.71	2.7
CPS	Mar 1998	2.7	2.7
CPS	Apr 1998	2.7	2.7
CPS	May 1998	2.69	2.69
CPS	Jun 1998	2.7	2.7
CPS	Jul 1998	2.71	2.7
CPS	Aug 1998	2.71	2.7
CPS	Sep 1998	2.71	2.71
CPS	Oct 1998	2.73	2.72
CPS	Nov 1998	2.73	2.72
CPS	Dec 1998	2.72	2.72
CPS	Jan 1999	2.72	2.72
CPS	Feb 1999	2.72	2.73
CPS	Mar 1999	2.73	2.73
CPS	Apr 1999	2.72	2.72
CPS	May 1999	2.72	2.72
CPS	Jun 1999	2.72	2.73
CPS	Jul 1999	2.72	2.73
CPS	Aug 1999	2.73	2.73
CPS	Sep 1999	2.73	2.73
CPS	Oct 1999	2.75	2.75
CPS	Nov 1999	2.73	2.74
CPS	Dec 1999	2.73	2.74
CPS	Jan 2000	2.74	2.74
CPS	Feb 2000	2.74	2.74
CPS	Mar 2000	2.74	2.74
CPS	Apr 2000	2.74	2.73
CPS	May 2000	2.73	2.74
CPS	Jun 2000	2.73	2.73
CPS	Jul 2000	2.74	2.74
CPS	Aug 2000	2.75	2.75
CPS	Sep 2000	2.74	2.75
CPS	Oct 2000	2.75	2.75
CPS	Nov 2000	2.74	2.74
CPS	Dec 2000	2.75	2.75
CPS	Jan 2001	2.75	2.75
CPS	Feb 2001	2.74	2.74
CPS	Mar 2001	2.75	2.76
CPS	Apr 2001	2.75	2.76
CPS	May 2001	2.75	2.76
CPS	Jun 2001	2.75	2.75
CPS	Jul 2001	2.75	2.75
CPS	Aug 2001	2.76	2.76
CPS	Sep 2001	2.75	2.76
CPS	Oct 2001	2.77	2.78
CPS	Nov 2001	2.77	2.77
CPS	Dec 2001	2.77	2.77
CPS	Jan 2002	2.78	2.78
CPS	Feb 2002	2.77	2.78
CPS	Mar 2002	2.77	2.77
CPS	Apr 2002	2.77	2.77
CPS	May 2002	2.76	2.77
CPS	Jun 2002	2.77	2.78
CPS	Jul 2002	2.76	2.77
CPS	Aug 2002	2.76	2.77
CPS	Sep 2002	2.77	2.77
CPS	Oct 2002	2.77	2.77
CPS	Nov 2002	2.78	2.79
CPS	Dec 2002	2.76	2.76
CPS	Jan 2003	2.76	2.77
CPS	Feb 2003	2.77	2.77
CPS	Mar 2003	2.77	2.77
CPS	Apr 2003	2.77	2.78
CPS	May 2003	2.77	2.77
CPS	Jun 2003	2.77	2.77
CPS	Jul 2003	2.76	2.77
CPS	Aug 2003	2.77	2.77
CPS	Sep 2003	2.77	2.77
CPS	Oct 2003	2.79	2.78
CPS	Nov 2003	2.77	2.78
CPS	Dec 2003	2.77	2.78
CPS	Jan 2004	2.77	2.78
CPS	Feb 2004	2.78	2.79
CPS	Mar 2004	2.77	2.78
CPS	Apr 2004	2.78	2.78
CPS	May 2004	2.76	2.77
CPS	Jun 2004	2.76	2.76
CPS	Jul 2004	2.76	2.76
CPS	Aug 2004	2.76	2.77
CPS	Sep 2004	2.77	2.78
CPS	Oct 2004	2.76	2.76
CPS	Nov 2004	2.76	2.77
CPS	Dec 2004	2.75	2.76
CPS	Jan 2005	2.78	2.79
CPS	Feb 2005	2.78	2.79
CPS	Mar 2005	2.76	2.77
CPS	Apr 2005	2.76	2.77
CPS	May 2005	2.76	2.76
CPS	Jun 2005	2.75	2.76
CPS	Jul 2005	2.74	2.75
CPS	Aug 2005	2.74	2.75
CPS	Sep 2005	2.75	2.75
CPS	Oct 2005	2.75	2.75
CPS	Nov 2005	2.75	2.75
CPS	Dec 2005	2.75	2.76
CPS	Jan 2006	2.76	2.77
CPS	Feb 2006	2.76	2.77
CPS	Mar 2006	2.76	2.77
CPS	Apr 2006	2.76	2.77
CPS	May 2006	2.75	2.75
CPS	Jun 2006	2.75	2.76
CPS	Jul 2006	2.75	2.76
CPS	Aug 2006	2.75	2.75
CPS	Sep 2006	2.75	2.76
CPS	Oct 2006	2.77	2.77
CPS	Nov 2006	2.77	2.77
CPS	Dec 2006	2.76	2.76
CPS	Jan 2007	2.77	2.77
CPS	Feb 2007	2.77	2.78
CPS	Mar 2007	2.78	2.78
CPS	Apr 2007	2.77	2.77
CPS	May 2007	2.75	2.76
CPS	Jun 2007	2.76	2.77
CPS	Jul 2007	2.77	2.77
CPS	Aug 2007	2.77	2.77
CPS	Sep 2007	2.76	2.77
CPS	Oct 2007	2.77	2.77
CPS	Nov 2007	2.76	2.77
CPS	Dec 2007	2.76	2.75
CPS	Jan 2008	2.77	2.77
CPS	Feb 2008	2.78	2.79
CPS	Mar 2008	2.78	2.78
CPS	Apr 2008	2.78	2.78
CPS	May 2008	2.77	2.78
CPS	Jun 2008	2.77	2.77
CPS	Jul 2008	2.75	2.76
CPS	Aug 2008	2.76	2.77
CPS	Sep 2008	2.77	2.77
CPS	Oct 2008	2.79	2.79
CPS	Nov 2008	2.79	2.79
CPS	Dec 2008	2.79	2.8
CPS	Jan 2009	2.8	2.81
CPS	Feb 2009	2.81	2.81
CPS	Mar 2009	2.79	2.79
CPS	Apr 2009	2.8	2.8
CPS	May 2009	2.8	2.81
CPS	Jun 2009	2.8	2.8
CPS	Jul 2009	2.8	2.79
CPS	Aug 2009	2.79	2.79
CPS	Sep 2009	2.8	2.8
CPS	Oct 2009	2.79	2.79
CPS	Nov 2009	2.8	2.81
CPS	Dec 2009	2.78	2.78
CPS	Jan 2010	2.81	2.81
CPS	Feb 2010	2.79	2.79
CPS	Mar 2010	2.79	2.78
CPS	Apr 2010	2.79	2.79
CPS	May 2010	2.79	2.79
CPS	Jun 2010	2.79	2.79
CPS	Jul 2010	2.78	2.78
CPS	Aug 2010	2.78	2.78
CPS	Sep 2010	2.8	2.8
CPS	Oct 2010	2.79	2.78
CPS	Nov 2010	2.8	2.8
CPS	Dec 2010	2.77	2.77

In figures 2 and 3, we show overall employment and mean wage levels in the CPS and OES (observations with imputed wages are included in employment totals, and observations from the incorporated self-employed are included in employment totals for the CPS even though no wage data are collected). The CPS data in figures 2 and 3 are monthly data, not seasonally adjusted. The OES data in these figures are from each data collection panel (the Oct/Nov/Dec surveys in 1998–2001, and each May and November survey in 2002–2010). All wage data are adjusted for inflation. The employment and mean wage series have similar trends in the two data sources, although the OES data show higher mean wages, particularly when imputed responses are included. This may be because establishments paying higher wages are more likely to respond to the survey, or because nonrespondents in smaller areas, which tend to have lower wages, are less likely to be imputed with respondents in the same area, and may have their wage structure imputed from respondents in other, higher wage areas.

Comparisons of variance trends in the CPS and OES data

The interval-based collection of wage data in the OES will tend to reduce the variance of wages calculated from the OES data relative to that calculated from CPS data. However, as explained above, each wage observation in the CPS can be assigned to a wage interval like those used in the OES, and each wage observation in the CPS can be assigned an intervalized wage corresponding to the mean wage for all CPS observations in the same interval at the same date. Thus, we can calculate the variance of wages in the CPS for both the “continuous” and the intervalized data. A time series for both these versions of the variance of ln(wage) from the CPS, as well as the variance of ln(wage) in the OES, is shown in figures 4 and 5.²⁴ As in figures 2 and 3 (and in all successive figures), the CPS data in figures 4 and 5 are monthly, and the OES data are from each survey panel. Although variance levels in the CPS are slightly lower when wages are intervalized, the time-series properties of these series are very similar. Both versions of ln(wage) variances are lower in the CPS than in the OES, although time trends from both datasets are similar. More curiously, dropping the imputed wage values has different impacts on measured wage variance trends in the CPS and the OES: dropping the imputed wages slightly increases wage variance in the CPS, but decreases wage variance in the OES.

View Chart Data

 

Figure 4. Overall log wage variance trends in the CPS and OES, 1998–2010

Data source	Date	Wage variance	Continuous wage variance
OES	Nov 1998	0.326	—
OES	Nov 1999	0.341	—
OES	Nov 2000	0.344	—
OES	Nov 2001	0.349	—
OES	Nov 2002	0.372	—
OES	May 2003	0.361	—
OES	Nov 2003	0.366	—
OES	May 2004	0.365	—
OES	Nov 2004	0.367	—
OES	May 2005	0.364	—
OES	Nov 2005	0.375	—
OES	May 2006	0.365	—
OES	Nov 2006	0.367	—
OES	May 2007	0.359	—
OES	Nov 2007	0.369	—
OES	May 2008	0.366	—
OES	Nov 2008	0.375	—
OES	May 2009	0.365	—
OES	Nov 2009	0.370	—
OES	May 2010	0.373	—
CPS	Jan 1998	0.2937	0.2978
CPS	Feb 1998	0.2881	0.2922
CPS	Mar 1998	0.2908	0.295
CPS	Apr 1998	0.2879	0.2921
CPS	May 1998	0.282	0.2862
CPS	Jun 1998	0.2863	0.2904
CPS	Jul 1998	0.2901	0.2942
CPS	Aug 1998	0.2879	0.2921
CPS	Sep 1998	0.2978	0.3019
CPS	Oct 1998	0.3006	0.3047
CPS	Nov 1998	0.2985	0.3027
CPS	Dec 1998	0.3	0.3041
CPS	Jan 1999	0.2879	0.2928
CPS	Feb 1999	0.2959	0.3009
CPS	Mar 1999	0.2881	0.2931
CPS	Apr 1999	0.2891	0.2943
CPS	May 1999	0.2982	0.3032
CPS	Jun 1999	0.2933	0.2983
CPS	Jul 1999	0.2968	0.3017
CPS	Aug 1999	0.2953	0.3002
CPS	Sep 1999	0.2931	0.298
CPS	Oct 1999	0.3056	0.3105
CPS	Nov 1999	0.2932	0.2982
CPS	Dec 1999	0.2986	0.3035
CPS	Jan 2000	0.2906	0.2957
CPS	Feb 2000	0.3038	0.3091
CPS	Mar 2000	0.2998	0.3049
CPS	Apr 2000	0.2943	0.2994
CPS	May 2000	0.3047	0.3096
CPS	Jun 2000	0.3064	0.3113
CPS	Jul 2000	0.3192	0.3243
CPS	Aug 2000	0.3061	0.3111
CPS	Sep 2000	0.2989	0.3039
CPS	Oct 2000	0.2985	0.3035
CPS	Nov 2000	0.2937	0.2988
CPS	Dec 2000	0.3069	0.312
CPS	Jan 2001	0.3065	0.3116
CPS	Feb 2001	0.308	0.3129
CPS	Mar 2001	0.3103	0.3154
CPS	Apr 2001	0.3027	0.3077
CPS	May 2001	0.2989	0.304
CPS	Jun 2001	0.3066	0.3117
CPS	Jul 2001	0.3065	0.3115
CPS	Aug 2001	0.3033	0.3083
CPS	Sep 2001	0.3058	0.3108
CPS	Oct 2001	0.3028	0.3077
CPS	Nov 2001	0.2947	0.2996
CPS	Dec 2001	0.3063	0.3112
CPS	Jan 2002	0.3092	0.3144
CPS	Feb 2002	0.3128	0.3179
CPS	Mar 2002	0.3129	0.3182
CPS	Apr 2002	0.3051	0.3101
CPS	May 2002	0.3023	0.3075
CPS	Jun 2002	0.3038	0.3089
CPS	Jul 2002	0.3102	0.3153
CPS	Aug 2002	0.3127	0.3179
CPS	Sep 2002	0.3055	0.3106
CPS	Oct 2002	0.3158	0.321
CPS	Nov 2002	0.3037	0.3088
CPS	Dec 2002	0.3038	0.3089
CPS	Jan 2003	0.3085	0.3136
CPS	Feb 2003	0.3171	0.3222
CPS	Mar 2003	0.3132	0.3183
CPS	Apr 2003	0.3123	0.3174
CPS	May 2003	0.3069	0.312
CPS	Jun 2003	0.3134	0.3185
CPS	Jul 2003	0.3151	0.3203
CPS	Aug 2003	0.3145	0.3196
CPS	Sep 2003	0.31	0.3151
CPS	Oct 2003	0.3194	0.3245
CPS	Nov 2003	0.3143	0.3194
CPS	Dec 2003	0.31	0.3152
CPS	Jan 2004	0.3165	0.3216
CPS	Feb 2004	0.3191	0.3242
CPS	Mar 2004	0.3143	0.3193
CPS	Apr 2004	0.3171	0.322
CPS	May 2004	0.3159	0.321
CPS	Jun 2004	0.3158	0.3207
CPS	Jul 2004	0.315	0.3201
CPS	Aug 2004	0.3171	0.3222
CPS	Sep 2004	0.3261	0.3312
CPS	Oct 2004	0.3172	0.3222
CPS	Nov 2004	0.3277	0.3327
CPS	Dec 2004	0.3214	0.3264
CPS	Jan 2005	0.3265	0.3313
CPS	Feb 2005	0.3357	0.3407
CPS	Mar 2005	0.3241	0.3291
CPS	Apr 2005	0.3192	0.3242
CPS	May 2005	0.3251	0.33
CPS	Jun 2005	0.3251	0.33
CPS	Jul 2005	0.3184	0.325
CPS	Aug 2005	0.318	0.3247
CPS	Sep 2005	0.3241	0.3305
CPS	Oct 2005	0.3239	0.3304
CPS	Nov 2005	0.3194	0.326
CPS	Dec 2005	0.3157	0.3223
CPS	Jan 2006	0.3203	0.3268
CPS	Feb 2006	0.3266	0.3333
CPS	Mar 2006	0.3222	0.3285
CPS	Apr 2006	0.3196	0.3258
CPS	May 2006	0.321	0.3272
CPS	Jun 2006	0.3266	0.333
CPS	Jul 2006	0.3228	0.3289
CPS	Aug 2006	0.3247	0.331
CPS	Sep 2006	0.3182	0.3244
CPS	Oct 2006	0.3223	0.3286
CPS	Nov 2006	0.3192	0.3253
CPS	Dec 2006	0.3246	0.331
CPS	Jan 2007	0.3126	0.3179
CPS	Feb 2007	0.3182	0.3235
CPS	Mar 2007	0.3224	0.3278
CPS	Apr 2007	0.3198	0.3251
CPS	May 2007	0.3128	0.318
CPS	Jun 2007	0.3238	0.329
CPS	Jul 2007	0.3284	0.3337
CPS	Aug 2007	0.3202	0.3253
CPS	Sep 2007	0.3119	0.3172
CPS	Oct 2007	0.3236	0.3289
CPS	Nov 2007	0.3169	0.3221
CPS	Dec 2007	0.3161	0.3213
CPS	Jan 2008	0.3094	0.3141
CPS	Feb 2008	0.3232	0.3279
CPS	Mar 2008	0.3276	0.3323
CPS	Apr 2008	0.3269	0.3316
CPS	May 2008	0.3342	0.3388
CPS	Jun 2008	0.3334	0.338
CPS	Jul 2008	0.3279	0.3341
CPS	Aug 2008	0.3356	0.3416
CPS	Sep 2008	0.343	0.3489
CPS	Oct 2008	0.3338	0.3398
CPS	Nov 2008	0.3353	0.3414
CPS	Dec 2008	0.3149	0.3209
CPS	Jan 2009	0.3186	0.3233
CPS	Feb 2009	0.3349	0.3397
CPS	Mar 2009	0.3145	0.3193
CPS	Apr 2009	0.3174	0.3222
CPS	May 2009	0.3228	0.3276
CPS	Jun 2009	0.3355	0.3402
CPS	Jul 2009	0.3385	0.3432
CPS	Aug 2009	0.3299	0.3345
CPS	Sep 2009	0.3381	0.3427
CPS	Oct 2009	0.3338	0.3385
CPS	Nov 2009	0.3284	0.333
CPS	Dec 2009	0.3264	0.331
CPS	Jan 2010	0.3489	0.3537
CPS	Feb 2010	0.3354	0.3401
CPS	Mar 2010	0.3312	0.3359
CPS	Apr 2010	0.3358	0.3404
CPS	May 2010	0.3311	0.3358
CPS	Jun 2010	0.3278	0.3324
CPS	Jul 2010	0.3398	0.3444
CPS	Aug 2010	0.338	0.3427
CPS	Sep 2010	0.3449	0.3495
CPS	Oct 2010	0.3407	0.3453
CPS	Nov 2010	0.3407	0.3453
CPS	Dec 2010	0.3352	0.34

View Chart Data

 

Figure 5. Log wage variance trends (no imputations) in the CPS and OES, 1998–2010

Data source	Date	Wage variance—no imputations	Continuous wage variance—no imputations
OES	Nov 1998	0.326	—
OES	Nov 1999	0.339	—
OES	Nov 2000	0.341	—
OES	Nov 2001	0.344	—
OES	Nov 2002	0.367	—
OES	May 2003	0.358	—
OES	Nov 2003	0.365	—
OES	May 2004	0.362	—
OES	Nov 2004	0.362	—
OES	May 2005	0.357	—
OES	Nov 2005	0.371	—
OES	May 2006	0.36	—
OES	Nov 2006	0.359	—
OES	May 2007	0.353	—
OES	Nov 2007	0.364	—
OES	May 2008	0.361	—
OES	Nov 2008	0.372	—
OES	May 2009	0.358	—
OES	Nov 2009	0.361	—
OES	May 2010	0.365	—
CPS	Jan 1998	0.295	0.2991
CPS	Feb 1998	0.2916	0.2958
CPS	Mar 1998	0.2923	0.2966
CPS	Apr 1998	0.2909	0.2951
CPS	May 1998	0.2833	0.2875
CPS	Jun 1998	0.2908	0.295
CPS	Jul 1998	0.2962	0.3004
CPS	Aug 1998	0.2903	0.2946
CPS	Sep 1998	0.3023	0.3064
CPS	Oct 1998	0.303	0.3071
CPS	Nov 1998	0.3015	0.3058
CPS	Dec 1998	0.3015	0.3057
CPS	Jan 1999	0.2913	0.2963
CPS	Feb 1999	0.3012	0.3062
CPS	Mar 1999	0.2885	0.2934
CPS	Apr 1999	0.29	0.2952
CPS	May 1999	0.303	0.308
CPS	Jun 1999	0.2976	0.3025
CPS	Jul 1999	0.3047	0.3096
CPS	Aug 1999	0.298	0.303
CPS	Sep 1999	0.3002	0.3051
CPS	Oct 1999	0.3048	0.3096
CPS	Nov 1999	0.2976	0.3026
CPS	Dec 1999	0.2978	0.3027
CPS	Jan 2000	0.2916	0.2967
CPS	Feb 2000	0.3096	0.3149
CPS	Mar 2000	0.3054	0.3106
CPS	Apr 2000	0.2971	0.3023
CPS	May 2000	0.3067	0.3116
CPS	Jun 2000	0.3099	0.3149
CPS	Jul 2000	0.3208	0.3259
CPS	Aug 2000	0.3112	0.3162
CPS	Sep 2000	0.3034	0.3084
CPS	Oct 2000	0.3025	0.3074
CPS	Nov 2000	0.3019	0.307
CPS	Dec 2000	0.3111	0.3162
CPS	Jan 2001	0.3085	0.3135
CPS	Feb 2001	0.3122	0.3172
CPS	Mar 2001	0.3201	0.3252
CPS	Apr 2001	0.3145	0.3195
CPS	May 2001	0.3061	0.3112
CPS	Jun 2001	0.3136	0.3185
CPS	Jul 2001	0.313	0.318
CPS	Aug 2001	0.3124	0.3174
CPS	Sep 2001	0.3106	0.3155
CPS	Oct 2001	0.3121	0.3171
CPS	Nov 2001	0.3016	0.3065
CPS	Dec 2001	0.3167	0.3216
CPS	Jan 2002	0.3128	0.3181
CPS	Feb 2002	0.3174	0.3225
CPS	Mar 2002	0.3218	0.3271
CPS	Apr 2002	0.3115	0.3165
CPS	May 2002	0.3088	0.314
CPS	Jun 2002	0.3127	0.3177
CPS	Jul 2002	0.318	0.3231
CPS	Aug 2002	0.3198	0.325
CPS	Sep 2002	0.3154	0.3203
CPS	Oct 2002	0.3233	0.3285
CPS	Nov 2002	0.3091	0.3141
CPS	Dec 2002	0.3083	0.3133
CPS	Jan 2003	0.3163	0.3214
CPS	Feb 2003	0.327	0.3321
CPS	Mar 2003	0.325	0.3301
CPS	Apr 2003	0.3228	0.3278
CPS	May 2003	0.3183	0.3233
CPS	Jun 2003	0.3282	0.3332
CPS	Jul 2003	0.327	0.3321
CPS	Aug 2003	0.3213	0.3264
CPS	Sep 2003	0.3193	0.3243
CPS	Oct 2003	0.3228	0.3279
CPS	Nov 2003	0.329	0.3341
CPS	Dec 2003	0.3148	0.3198
CPS	Jan 2004	0.3287	0.3336
CPS	Feb 2004	0.331	0.336
CPS	Mar 2004	0.3265	0.3314
CPS	Apr 2004	0.3282	0.3331
CPS	May 2004	0.3184	0.3234
CPS	Jun 2004	0.3286	0.3336
CPS	Jul 2004	0.3208	0.3258
CPS	Aug 2004	0.3261	0.3311
CPS	Sep 2004	0.3362	0.3412
CPS	Oct 2004	0.3223	0.3272
CPS	Nov 2004	0.3319	0.3368
CPS	Dec 2004	0.3288	0.3337
CPS	Jan 2005	0.3369	0.3418
CPS	Feb 2005	0.3407	0.3456
CPS	Mar 2005	0.3372	0.3422
CPS	Apr 2005	0.3293	0.3343
CPS	May 2005	0.3366	0.3414
CPS	Jun 2005	0.3374	0.3423
CPS	Jul 2005	0.3239	0.3307
CPS	Aug 2005	0.3259	0.3327
CPS	Sep 2005	0.3331	0.3395
CPS	Oct 2005	0.3319	0.3385
CPS	Nov 2005	0.3255	0.3324
CPS	Dec 2005	0.3283	0.335
CPS	Jan 2006	0.3314	0.338
CPS	Feb 2006	0.3361	0.3426
CPS	Mar 2006	0.3355	0.342
CPS	Apr 2006	0.3286	0.3349
CPS	May 2006	0.3309	0.3372
CPS	Jun 2006	0.3393	0.3458
CPS	Jul 2006	0.3276	0.3338
CPS	Aug 2006	0.3379	0.3444
CPS	Sep 2006	0.3275	0.3337
CPS	Oct 2006	0.3297	0.3361
CPS	Nov 2006	0.3259	0.3321
CPS	Dec 2006	0.3317	0.3381
CPS	Jan 2007	0.3201	0.3255
CPS	Feb 2007	0.3296	0.3349
CPS	Mar 2007	0.3294	0.3348
CPS	Apr 2007	0.3226	0.3278
CPS	May 2007	0.3228	0.3281
CPS	Jun 2007	0.333	0.3382
CPS	Jul 2007	0.3276	0.333
CPS	Aug 2007	0.3266	0.3318
CPS	Sep 2007	0.3231	0.3284
CPS	Oct 2007	0.3277	0.333
CPS	Nov 2007	0.3232	0.3284
CPS	Dec 2007	0.3194	0.3247
CPS	Jan 2008	0.3197	0.3242
CPS	Feb 2008	0.3286	0.3333
CPS	Mar 2008	0.3298	0.3344
CPS	Apr 2008	0.3378	0.3424
CPS	May 2008	0.3401	0.3446
CPS	Jun 2008	0.3464	0.351
CPS	Jul 2008	0.3356	0.3419
CPS	Aug 2008	0.3396	0.3457
CPS	Sep 2008	0.351	0.3571
CPS	Oct 2008	0.3413	0.3476
CPS	Nov 2008	0.3411	0.3474
CPS	Dec 2008	0.3171	0.3231
CPS	Jan 2009	0.3252	0.33
CPS	Feb 2009	0.3401	0.3449
CPS	Mar 2009	0.3207	0.3255
CPS	Apr 2009	0.327	0.3318
CPS	May 2009	0.3291	0.3339
CPS	Jun 2009	0.344	0.3488
CPS	Jul 2009	0.3459	0.3507
CPS	Aug 2009	0.3319	0.3366
CPS	Sep 2009	0.3474	0.3521
CPS	Oct 2009	0.3404	0.3451
CPS	Nov 2009	0.3382	0.3428
CPS	Dec 2009	0.3335	0.3383
CPS	Jan 2010	0.3526	0.3574
CPS	Feb 2010	0.3453	0.35
CPS	Mar 2010	0.3372	0.342
CPS	Apr 2010	0.3419	0.3465
CPS	May 2010	0.3405	0.3453
CPS	Jun 2010	0.3377	0.3424
CPS	Jul 2010	0.3455	0.3501
CPS	Aug 2010	0.3443	0.3491
CPS	Sep 2010	0.3516	0.3563
CPS	Oct 2010	0.3426	0.3473
CPS	Nov 2010	0.3505	0.3552
CPS	Dec 2010	0.3423	0.3471

We can also compare ln(wage) variance trends in the CPS and OES within sectors, industries, and occupations. Sector is defined as private industry, federal government, state government, and local government. Variance trends in the CPS and OES—dropping imputed values from both datasets—are shown by sector in figure 6, by industry in figure 7, and by occupation in figure 8.

In figure 6, we show the trends in ln(wage) variances by sector. Although the OES and CPS show similar patterns of ln(wage) variance by sector and over time, the higher wage variance in the OES data than in the CPS data is driven by the wage variance within the private sector. For federal government workers, ln(wage) variance is higher in the CPS than the OES. Also of interest in figure 6 is that both data sources show that wage variance has increased over time in the federal government sector and in the private sector, but not in the state and local government sectors.